def find_nodes(longitude, latitude, grid_of_nodes, node_info, n):
if out_of_bounds(longitude, latitude, node_info):
print "OUT OF BOUNDS: (Long, Lat) = (" + str(longitude) + ", " + (
str(latitude) + ")")
return None
# Node closest to coords and its distance
best_node = None
best_distance = 1000000
i = int(float(longitude - node_info[3]) *
(n / float(node_info[2] - node_info[3])))
j = int(float(latitude - node_info[1]) *
(n / float(node_info[0] - node_info[1])))
# You have to check the surrounding area (if a coordinate is really close
# to the edge of the region[i][j] it could be in a different region
# [i - 1][j] for example
if i != 0:
i -= 1
if j != 0:
j -= 1
for n in range(3):
if i + n >= len(grid_of_nodes):
break
for m in range(3):
if j + m >= len(grid_of_nodes[0]):
break
grid_region = grid_of_nodes[i + n][j + m]
for node in grid_region.nodes:
currDistance = distance(latitude, longitude, node.lat,
node.long)
if currDistance < best_distance and (
len(node.forkward_links) > 0):
best_node = node
best_distance = currDistance
return best_node
def find_nodes(longitude, latitude, grid_of_nodes, node_info, n):
if out_of_bounds(longitude, latitude, node_info):
print "OUT OF BOUNDS: (Long, Lat) = (" + (str(longitude) + ", " +
str(latitude) + ")")
return None
# node closest to coords and its distance
best_node = None
best_dist = 1000000
i = int(
float(longitude - node_info[3]) *
(n / float(node_info[2] - node_info[3])))
j = int(
float(latitude - node_info[1]) *
(n / float(node_info[0] - node_info[1])))
# You have to check the surrounding area (if a coordinate is really close
# to the edge of the region[i][j] it could be in a different region
# [i - 1][j] for example
if i != 0:
i -= 1
if j != 0:
j -= 1
for n in range(3):
if i + n >= len(grid_of_nodes):
break
for m in range(3):
if j + m >= len(grid_of_nodes[0]):
break
grid_region = grid_of_nodes[i + n][j + m]
for node in grid_region.nodes:
curr_dist = distance(latitude, longitude, node.lat, node.long)
if curr_dist < best_dist and (len(node.forward_links) > 0):
best_node = node
best_dist = curr_dist
return best_node
def PickupRequest(self, TAXINOTNEEDED, arrayFromCSV, timeRequested):
global searchAlgorithm, searchRadius
#Occasionally a customer's data is bad, and we opt to throw those events out
if float(arrayFromCSV[10]) == 0 or float(arrayFromCSV[11]) == 0 or float(arrayFromCSV[12]) == 0 or float(arrayFromCSV[13]) == 0:
print "BAD CUSTOMER longitude/latitude"
return
#If there are no taxis available, we assign the customers to the waiting array so when a taxi is freed up we can pick them up
if len(self.taxisAvailable) == 0:
self.customersWaiting.append((timeRequested, arrayFromCSV))
self.taxiNeeded += 1
return
custLong = float(arrayFromCSV[10])
custLat = float(arrayFromCSV[11])
taxiList = []
#We sort the taxis by their distance by making tuples (distance, taxi)
for taxi in self.taxisAvailable:
taxiList.append((distance(taxi.lat, taxi.long, custLat, custLong), taxi))
taxiList.sort(key = lambda taxi: taxi[0])
#Using either the A* algorithm ('A') or the winding factor ('W') finds the closest taxi.
bestTaxi = None
best_distance = 10000000
if searchAlgorithm == 'A':
for taxi in taxiList:
#We sorted them all by Euclidean distance - if the current best distance is smaller than the current euclidean distance (the shortest distance) then we don't need to check anymore as none can be closer
if taxi[0] > best_distance or taxi[0] > searchRadius:
break
currDistance = AStar(taxi[1].long, taxi[1].lat, custLong, custLat, self.grid, self.node_info, self.gridSize)
if currDistance < best_distance:
bestTaxi = taxi[1]
best_distance = currDistance
if searchAlgorithm == 'W':
for taxi in taxiList:
#We sorted them all by Euclidean distance - if the current best distance is smaller than the current euclidean distance (the shortest distance) then we don't need to check anymore as none can be closer
if taxi[0] > best_distance or taxi[0] > searchRadius:
break
currDistance = estimateActual(taxi[1].lat, taxi[1].long, custLat, custLong)
if currDistance < best_distance:
bestTaxi = taxi[1]
best_distance = currDistance
#If all were outside the search radius, then no taxis will have picked them up and thus no customers would have been found
if bestTaxi is None:
self.customersWaiting.append((timeRequested, arrayFromCSV))
self.taxiNeeded += 1
return
#This is what we are hoping to minimize, so we keep track of it all
self.distanceTravelled += best_distance
#For now, convert to time by assuming they are going twenty miles per hour
bestTime = (best_distance / float(5280)) / 20 * 3600 #20 mph and 3600 seconds per hour
#We must remove the taxi from available taxis but add them to the ones in use
self.taxisInUse.add(bestTaxi)
self.taxisAvailable.discard(bestTaxi)
self.generatePickup(bestTaxi, arrayFromCSV, changeTime(timeRequested, bestTime))
def PickupRequest(self, TAXINOTNEEDED, arrayFromCSV, timeRequested):
global searchAlgorithm, searchRadius
#Occasionally a customer's data is bad, and we opt to throw those events out
if float(arrayFromCSV[10]) == 0 or float(
arrayFromCSV[11]) == 0 or float(
arrayFromCSV[12]) == 0 or float(arrayFromCSV[13]) == 0:
print "BAD CUSTOMER longitude/latitude"
return
#If there are no taxis available, we assign the customers to the waiting array so when a taxi is freed up we can pick them up
if len(self.taxisAvailable) == 0:
self.customersWaiting.append((timeRequested, arrayFromCSV))
self.taxiNeeded += 1
return
custLong = float(arrayFromCSV[10])
custLat = float(arrayFromCSV[11])
taxiList = []
#We sort the taxis by their distance by making tuples (distance, taxi)
for taxi in self.taxisAvailable:
taxiList.append((distance(taxi.lat, taxi.long, custLat,
custLong), taxi))
taxiList.sort(key=lambda taxi: taxi[0])
#Using either the A* algorithm ('A') or the winding factor ('W') finds the closest taxi.
bestTaxi = None
best_distance = 10000000
if searchAlgorithm == 'A':
for taxi in taxiList:
#We sorted them all by Euclidean distance - if the current best distance is smaller than the current euclidean distance (the shortest distance) then we don't need to check anymore as none can be closer
if taxi[0] > best_distance or taxi[0] > searchRadius:
break
currDistance = AStar(taxi[1].long, taxi[1].lat, custLong,
custLat, self.grid, self.node_info,
self.gridSize)
if currDistance < best_distance:
bestTaxi = taxi[1]
best_distance = currDistance
if searchAlgorithm == 'W':
for taxi in taxiList:
#We sorted them all by Euclidean distance - if the current best distance is smaller than the current euclidean distance (the shortest distance) then we don't need to check anymore as none can be closer
if taxi[0] > best_distance or taxi[0] > searchRadius:
break
currDistance = estimateActual(taxi[1].lat, taxi[1].long,
custLat, custLong)
if currDistance < best_distance:
bestTaxi = taxi[1]
best_distance = currDistance
#If all were outside the search radius, then no taxis will have picked them up and thus no customers would have been found
if bestTaxi is None:
self.customersWaiting.append((timeRequested, arrayFromCSV))
self.taxiNeeded += 1
return
#This is what we are hoping to minimize, so we keep track of it all
self.distanceTravelled += best_distance
#For now, convert to time by assuming they are going twenty miles per hour
bestTime = (best_distance /
float(5280)) / 20 * 3600 #20 mph and 3600 seconds per hour
#We must remove the taxi from available taxis but add them to the ones in use
self.taxisInUse.add(bestTaxi)
self.taxisAvailable.discard(bestTaxi)
self.generatePickup(bestTaxi, arrayFromCSV,
changeTime(timeRequested, bestTime))
def heuristic(node, end_node, max_speed):
return distance(node.lat, node.long, end_node.lat, end_node.long) / (
max_speed)
def heuristic(node, end_node, max_speed):
return distance(node.lat, node.long, end_node.lat,
end_node.long) / max_speed